In the cloud platform load balancing process, multiple Virtual Private Network (VPN) routing forwarding tables are required to isolate carriers in order to achieve multi-carrier access. Each VPN routing forwarding table needs to be separately deployed with a virtual load balancer to provide load balancing services for the carrier's traffic data. If the cloud platform has m tenants and each tenant needs to access n carriers, then m*n virtual load balancers need to be deploy to provide load balancing services. At this time, the number of virtual load balancers is positively related to the number of carriers, resulting in too many virtual load balancers required for the cloud platform, and low resource utilization of the virtual load balancer.